The goal of this work is to develop an explanatory and predictive model of eye movement planning in normal and low vision that can be used to objectively design and measure the effectiveness of low vision rehabilitation. In the previous project period, we developed a quantitative measure of eye movement efficiency and found that inefficiencies among maculopathy patients could not be explained simply by the extent of their central field loss. Thus, there must be additional factors that impair the eye movement strategies in these patients. In this proposal, we further develop our computational model to predict human saccade behavior in addition to fixation placement. The first two aims will empirically measure how and whether motor error limits saccade amplitude and direction. The third aim seeks to improve upon previous endeavors to establish a natural causal link between eye movement generation and stimulus uncertainty. The final aim applies the revised model to probe a possible source of eye movement inefficiency in maculopathy patients: we test the hypothesis that patients retain a "normal" internal template of their vision instead of a veridical representation of their central field loss. This hypothesis is consistent with clinical observations that patients'typically begin complaining about their vision only when bilateral scotomas exceed 10deg, and that teaching "scotoma awareness" can improve functional outcome. Our computational model will continue to provide a conceptual framework for challenging clinical hypotheses, with the eventual goal of transforming low vision rehabilitation practice into rehabilitation science.